Covariant transverse-traceless projection for secondary gravitational waves

TL;DR

This paper introduces a gauge-invariant method for analyzing second-order gravitational waves using a covariant transverse-traceless projection, resolving gauge dependence issues in previous perturbative approaches.

Contribution

It proposes a new covariant, gauge-invariant energy density for gravitational waves based on spacetime tensors, applicable across different gauges.

Findings

The new energy density contains only propagating modes in the Newtonian gauge.

It can be transformed to the synchronous gauge with consistent results.

The method avoids gauge dependence issues in second-order gravitational wave analysis.

Abstract

Second-order tensor modes induced by nonlinear gravity are a key component of the cosmological background of gravitational waves. A detection of this background would allow us to probe the primordial power spectrum at otherwise inaccessible scales. Usually, the energy density of these gravitational waves is studied within perturbation theory in a particular gauge -- a connection between our physical spacetime and a fictitious background. It is a widely recognized issue that the second-order, scalar-induced gravitational waves are gauge dependent. This issue arises because they are not well-defined as tensors in the physical spacetime at second-order and are thus unphysical. In this paper, we propose the covariant transverse-traceless projection of the extrinsic curvature to study cosmological gravitational waves on a spatial hypersurface. We define a new energy density which is based purely on spacetime tensors, independent of perturbation theory, and thus is gauge invariant by definition. We show that, in the context of second-order perturbation theory, this new energy density contains only propagating modes in the constant-time hypersurface in the Newtonian gauge. We further show that we can recover the same gravitational waves after a transformation to the synchronous gauge, so long as we correctly identify the Newtonian hypersurface.